MCP server for AI coding agents

RepoRadar runs a Model Context Protocol server that lets Claude Code, Codex, Cursor, or any MCP-compatible agent query tracked GitHub repository data directly — search, trending lists, category browsing, AI-generated summaries, and reports — without scraping pages or leaving the terminal.

Add it to your agent

If your client supports remote MCP servers over Streamable HTTP, point it directly at the endpoint:

{
  "mcpServers": {
    "reporadar": {
      "url": "https://reporadar.spreix.com/api/mcp"
    }
  }
}

Claude Code: claude mcp add --transport http reporadar https://reporadar.spreix.com/api/mcp

For stdio-only clients that don't yet support remote servers, use mcp-remote as a bridge:

{
  "mcpServers": {
    "reporadar": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://reporadar.spreix.com/api/mcp"]
    }
  }
}

Available tools

• search_repos — search tracked repositories by name, topic, or keyword
• get_repo — full profile for one repo: description, stars, AI summary, use cases, getting-started
• get_trending — what was trending on GitHub for a given day (defaults to latest)
• browse_category — list repos in a category, or list all categories
• get_report — a published daily, weekly, or monthly report

All five are read-only and backed by the same data that powers the site itself — nothing is agent-specific or served differently. See how the data is collected for details on sourcing and update cadence.

Why this exists

An agent helping someone evaluate or discover open-source tools shouldn't have to guess star counts or re-derive a project's use cases from a README on the fly —RepoRadar already tracks that, twice a day, across 1,058 repositories. This makes that data queryable directly instead of requiring a web fetch and re-parsing of the page.